public static void main(String[] args) {
Map<String, String> array = new HashMap<String, String>();
array.put("male", "M");
array.put("female", "F");
MNBTClassifer classifier = new MNBTClassifer(array);
Tokenizer tokenizer = new Tokenizer();
TokenizeDirectory dir = new TokenizeDirectory();
dir.DirectoryTokenizer("BLOGS", classifier.getClasses());
classifier.updatevocsize();
String dirktest = "TESTTEST.txt";
// String spam = "TESTMAILS/spmsgc34.txt";
// String ham = "TESTMAILS/8-817msg1.txt";
// System.out.println("V_SIZE " + classifier.getClasses()[0].getname() + "= " +
// classifier.getClasses()[0].getvocsize());
// System.out.println("V_SIZE " + classifier.getClasses()[1].getname() + "= " +
// classifier.getClasses()[1].getvocsize());
// System.out.println("wordprob /'company/' :" + classifier.wordprob("company",
// classifier.getClasses()[0]) + " class: " + classifier.getClasses()[0].getname());
// System.out.println("wordprob /'company/' :" + classifier.wordprob("company",
// classifier.getClasses()[1]) + " class: " + classifier.getClasses()[1].getname());
System.out.println(
"Supposed to be female: " + classifier.classify(tokenizer.tokenize(dirktest)));
// System.out.println("Supposed to be ham: " + classifier.classify(tokenizer.tokenize(ham)));
// TODO Auto-generated method stub
}
public static void main(String[] args) {
Map<String, String> array = new HashMap<String, String>();
array.put("spam", "spm");
array.put("ham", "msg");
MNBTClassifer classifier = new MNBTClassifer(array);
Tokenizer tokenizer = new Tokenizer();
TokenizeDirectory dir = new TokenizeDirectory();
dir.DirectoryTokenizer("EMAILS", classifier.getClasses());
classifier.updatevocsize();
String directorypath = "TESTMAILS";
ClassifierClass[] classes = classifier.getClasses();
TokenizeDirectory tokenizedirectory = new TokenizeDirectory();
double correct = 0;
double incorrect = 0;
File directory = new File(directorypath);
File[] directoryListing = directory.listFiles();
System.out.println(Arrays.toString(directoryListing));
if (directoryListing != null) {
for (File child : directoryListing) {
if (child.getName().endsWith(".txt")) {
String filename = directorypath + "/" + child.getName();
String category = tokenizedirectory.getCategory(classes, filename);
String totest = classifier.classify(tokenizer.tokenize(filename));
if (totest.equals(category)) {
correct += 1;
} else {
incorrect += 1;
}
System.out.println("supposed to be " + category + ": " + totest);
}
}
}
double percentage = correct / (correct + incorrect);
System.out.println("percentage correct: " + percentage);
}